Intelligent automation has helped boost medical studies. Researchers have developed techniques to help with cancer (melanoma) detection and prognosis since medical data has become freely accessible. In these circumstances, models generated by machine learning and deep learning offer a reliable, useful, and speedy solution. PRISMA parameters were used to choose literature from 2018 to 2023 that were listed in Web of Science, EMBASE, and EBSCO. To find scholarly manuscripts related to the present research that utilized AI-based acquiring methodologies for cancer evaluation, a successful search strategy was employed. Overall, 130 researches indicate that deep learning-based processors and traditional machine learning-based categories have a considerable impact on tumor prognosis. All things considered, the evaluation highlighted the faults of the earlier language and looked at the efforts of unique researchers. The results were compared using a variety of criteria, such as forecasting number, fact, empathy, accuracy, roll rating, identification proportion, zone addressed, preciseness, recall, and F1-score. The responses to the five scheduled investigations have been reviewed. Although several of the techniques suggested in the article have high accuracy for forecasting, mortality from cancer continues to rise. Therefore, further research is required to address the issues with tumor predicting.